The paper develops a procedure for detecting anomalies in geomagnetic field observations by digitally processing information from unmanned aerial vehicles during geomagnetic surveys. Data from a global navigation satellite system and from a scalar magnetometer were used. The geomagnetic anomaly detection was based on the mathematical apparatus of a system of approximation models using two-stage approximations. The first stage involved constructing a system of moving local approximation models that were defined on sequences of intersecting local subareas. An algorithm was formed to optimize the parameters of moving local approximation models based on solving systems of linear equations. The second stage was spatial filtering of the resulting system of local approximation models using weighting functions. The geomagnetic anomalies were detected under approaches from decision theory. The procedure that we developed for detecting geomagnetic anomalies was investigated from model observations using statistical testing and optimization search methods. Calculations of the probabilities of correct and false detection of geomagnetic anomalies are proposed. The method of the system of approximation models for experimental observations has been investigated, and the characteristics of geomagnetic anomaly detection have been evaluated. Satisfactory efficiency of the proposed detection procedure for geomagnetic anomalies was confirmed.
Getmanov et al. (Mon,) studied this question.